US11151130B2 - Systems and methods for assessing quality of input text using recurrent neural networks - Google Patents
Systems and methods for assessing quality of input text using recurrent neural networks Download PDFInfo
- Publication number
- US11151130B2 US11151130B2 US15/793,281 US201715793281A US11151130B2 US 11151130 B2 US11151130 B2 US 11151130B2 US 201715793281 A US201715793281 A US 201715793281A US 11151130 B2 US11151130 B2 US 11151130B2
- Authority
- US
- United States
- Prior art keywords
- input text
- sentences
- word
- incorrect
- text
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active, expires
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/242—Query formulation
- G06F16/243—Natural language query formulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/232—Orthographic correction, e.g. spell checking or vowelisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/242—Query formulation
- G06F16/2433—Query languages
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/242—Query formulation
- G06F16/2433—Query languages
- G06F16/2448—Query languages for particular applications; for extensibility, e.g. user defined types
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2453—Query optimisation
- G06F16/24534—Query rewriting; Transformation
- G06F16/2454—Optimisation of common expressions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2455—Query execution
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/3332—Query translation
- G06F16/3334—Selection or weighting of terms from queries, including natural language queries
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/35—Clustering; Classification
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/93—Document management systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/253—Grammatical analysis; Style critique
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2457—Query processing with adaptation to user needs
- G06F16/24575—Query processing with adaptation to user needs using context
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
Definitions
- the disclosure herein generally relate to assessment systems for input text, and, more particularly, to systems and methods for assessing quality of input text using recurrent neural networks.
- Embodiments of the present disclosure present technological improvements as solutions to one or more of the above-mentioned technical problems recognized by the inventors in conventional systems. For example, in one aspect, a processor implemented method for assessing quality of input text using recurrent neural networks is provided.
- the method comprises obtaining, via one or more hardware processors, an input text from a user, wherein the input text comprises a plurality of sentences, each sentence having a plurality of words; performing a comparison of each word from the input text with a dictionary stored in a database to determine a closest recommended word for each word in the input text; analyzing the input text to determine context of each word based on at least a portion of the input text; determining, based on the determined context, at least one of one or more correct sentences, one or more incorrect sentences, and one or more complex sentences from the input text; converting, using the determined context and at least one of the set of correct sentences and the set of incorrect sentences, and the one or more complex sentences, each word from the input text to a vector based on one or more concepts by comparing each word across the plurality of sentences of the input text to generate a set of vectors; and assessing, using one or more recurrent neural networks, quality of the input text based on at least one of the set of generated vectors, the comparison, the determined context and
- the step of assessing quality of the input text may comprises: assigning a weightage to each of the set of generated vectors, the comparison, the determined context and the at least one of the one or more correct sentences, the one or more incorrect sentences and the one or more complex sentences; and generating a score for the input text based on the weightage.
- a completeness score may be generated for the input text based on a topic associated with the input text.
- a system for assessing quality of input text using recurrent neural networks comprises a memory storing instructions; one or more communication interfaces; and one or more hardware processors coupled to said memory using said one or more communication interfaces, wherein said one or more hardware processors are configured by said instructions to: obtain an input text from a user, wherein the input text comprises a plurality of sentences, each sentence having a plurality of words, perform a comparison of each word from the input text with a dictionary stored in a database to determine a closest recommended word for each word in the input text, analyze the input text to determine context of each word based on at least a portion of the input text, determine, based on the determined context, at least one of one or more correct sentences, one or more incorrect sentences, and one or more complex sentences from the input text, convert, using the determined context and at least one of the set of correct sentences and the set of incorrect sentences, and the one or more complex sentences, each word from the input text to a vector based on one or more concepts by comparing each word
- the one or more hardware processors are configured to assess the quality of the input text by: assigning a weightage to each of the set of generated vectors, the comparison, the determined context and the at least one of the one or more correct sentences, the one or more incorrect sentences and the one or more complex sentences; and generating a score for the input text based on the weightage.
- the one or more hardware processors may be further configured to generate a completeness score for the input text based on a topic associated with the input text.
- one or more non-transitory machine readable information storage mediums comprising one or more instructions.
- the one or more instructions which when executed by one or more hardware processors causes obtaining an input text from a user, wherein the input text comprises a plurality of sentences, each sentence having a plurality of words; performing a comparison of each word from the input text with a dictionary stored in a database to determine a closest recommended word for each word in the input text; analyzing the input text to determine context of each word based on at least a portion of the input text; determining, based on the determined context, at least one of one or more correct sentences, one or more incorrect sentences, and one or more complex sentences from the input text; converting, using the determined context and at least one of the set of correct sentences and the set of incorrect sentences, and the one or more complex sentences, each word from the input text to a vector based on one or more concepts by comparing each word across the plurality of sentences of the input text to generate a set of vectors; and assessing, using one or more instructions
- the step of assessing quality of the input text may comprises: assigning a weightage to each of the set of generated vectors, the comparison, the determined context and the at least one of the one or more correct sentences, the one or more incorrect sentences and the one or more complex sentences; and generating a score for the input text based on the weightage.
- a completeness score may be generated for the input text based on a topic associated with the input text.
- FIG. 1 illustrates an exemplary block diagram of a system for assessing input text using recurrent neural networks in accordance with an embodiment of the present disclosure.
- FIG. 2 illustrates an exemplary flow diagram of a method for assessing input text using recurrent neural networks implemented with the system of FIG. 1 in accordance with an embodiment of the present disclosure.
- Input text assessment automation typically involves manually grade every written text provided by a user. This process is cumbersome and tedious. Every parameter to be measured is evaluated in a single reading and hence the quantitative assessment may not be precise.
- the embodiments of the present disclosure evaluates (or assesses) an input (written) text based on its content, clarity, spelling, grammar correction and coherence of text. Since these parameters are highly subjective in nature and therefore quantitatively measuring these is an acute problem. The system cannot be trained for measuring these as the rating of these may vary or change from evaluator (or reviewer) to evaluator, making it difficult to find the pattern in the marking.
- the embodiments of the present disclosure have presented experimental results wherein data has been scraped and used to train the system to develop a statistical model corresponding to a language (or languages). This enables the system of the present disclosure to be trained and assess quality of subsequent input text.
- the system implements a spell check module that if a given non-noun word when not present in the dictionary, checks for all words with one unit distance from the given word. Word with maximum proximity is selected. If more than one word has same proximity, then the one which occurs in a given context is chosen to be closest recommended word.
- the system extends this technique for higher level spell checks with multiple errors.
- the system implements a module to determine correct, incorrect and complex sentences that may have grammatical errors. For determining such sentences, the system implements one or more Application Programming Interface(s) (APIs) to cross check (or cross validate) with output that is being generated.
- APIs Application Programming Interface
- the system further performs a coherence check which measures inter sentence and intra sentence coherence.
- the coherence technique is achieved by training a neural network, for example, recurrent neural network, such as but not limited to a Long Short Term Memory (LSTM) network that learns to construct a sentence by selecting every word after another.
- a neural network for example, recurrent neural network, such as but not limited to a Long Short Term Memory (LSTM) network that learns to construct a sentence by selecting every word after another.
- LSTM Long Short Term Memory
- FIGS. 1 through 2 where similar reference characters denote corresponding features consistently throughout the figures, there are shown preferred embodiments and these embodiments are described in the context of the following exemplary system and/or method.
- FIG. 1 illustrates an exemplary block diagram of a system 100 for assessing input text according to an embodiment of the present disclosure.
- the system 100 includes one or more processors 104 , communication interface device(s) or input/output (I/O) interface(s) 106 , and one or more data storage devices or memory 102 operatively coupled to the one or more processors 104 .
- the one or more processors 104 that are hardware processors can be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, and/or any devices that manipulate signals based on operational instructions.
- the processor(s) is configured to fetch and execute computer-readable instructions stored in the memory.
- the system 100 can be implemented in a variety of computing systems, such as laptop computers, notebooks, hand-held devices, workstations, mainframe computers, servers, a network cloud and the like.
- the I/O interface device(s) 106 can include a variety of software and hardware interfaces, for example, a web interface, a graphical user interface, and the like and can facilitate multiple communications within a wide variety of networks N/W and protocol types, including wired networks, for example, LAN, cable, etc., and wireless networks, such as WLAN, cellular, or satellite.
- the I/O interface device(s) can include one or more ports for connecting a number of devices to one another or to another server.
- the memory 102 may include any computer-readable medium known in the art including, for example, volatile memory, such as static random access memory (SRAM) and dynamic random access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes.
- volatile memory such as static random access memory (SRAM) and dynamic random access memory (DRAM)
- non-volatile memory such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes.
- ROM read only memory
- erasable programmable ROM erasable programmable ROM
- FIG. 2 illustrates an exemplary flow diagram of a method for assessing input text using the system 100 of FIG. 1 in accordance with an embodiment of the present disclosure.
- the system 100 comprises one or more data storage devices or the memory 102 operatively coupled to the one or more hardware processors 104 and is configured to store instructions for execution of steps of the method by the one or more processors 104 .
- the steps of the method of the present disclosure will now be explained with reference to the components of the system 100 as depicted in FIG. 1 , and the flow diagram.
- the one or more processors 104 obtain an input text from a user.
- the input text comprises a plurality of sentences, wherein each sentence may include a plurality of words.
- the input text may be a typed text input, or handwritten document.
- the system 100 may perform an optical character recognition technique on the handwritten document and extract the input text from the same.
- the one or more hardware processors 104 perform a comparison of each word from the input text with a dictionary stored in a database to determine a closest recommended word for each word in the input text.
- a spell checker may be implemented to perform spell check of the input text wherein closest recommended word(s) may be provided for a given word in the input text by way of insertion, deletion, substitution or rotation of letters in a word.
- the dictionary may be stored in a remote system or a cloud computing environment, wherein the system 100 may query the remote system or the cloud computing environment and the comparison of each word from the input text with words from dictionary may be performed.
- the one or more hardware processors 104 analyze the input text to determine context of each word based on at least a portion of the input text.
- the system 100 employs a windowing technique that is applied on one or more select portion(s) of the input text to perform context analysis and determine context of each word. For example, assuming a paragraph having text of 4 sentences. For each word in a first sentence, windowing technique may be applied on the second and third sentence to determine the context of each word in the input text (or the first sentence).
- the one or more hardware processors 104 determine, based on the determined context, at least one of one or more correct sentences, one or more incorrect sentences, and/or one or more complex sentences from the input text.
- the system 100 may implement a grammar checker that uses one or more Application programming interface (APIs) to give one or more sentences with corrected grammar.
- APIs Application programming interface
- the system 100 may determine a count of corrections made and penalize the user with an increase in count. Depending on the severity of error made the penalty can be changed.
- the one or more hardware processors 104 convert each word from the input text to a vector based on one or more concepts by comparing each word across the plurality of sentences of the input text to generate a set of vectors.
- the above technique of converting each word from the input text to a vector is implemented to determine coherence of words and sentences in the input text. Every word is converted into a vector using the contexts the word appeared.
- These vectors with ‘N’ dimensions may be used to train the system 100 that implements one or more recurrent neural networks (RNNs).
- the recurrent neural network comprises a Long Short Term Memory network (e.g., LSTM network).
- LSTM network architecture comprising following expressions (or equations) that are used by the system 100 to learn the input text pattern and generate assessment report which determines quality of the input text:
- f t ⁇ g ( W f x t +U f h t-1 +b f )
- i t ⁇ g ( W i x t +U i h t-1 +b i )
- o t ⁇ g ( W o x t +U o h t-1 +b o )
- c t f t o c t-1 +i t o ⁇ c ( W c x t +U c h t-1 +b c )
- h t o t o ⁇ h ( c t )
- ⁇ g The original is a sigmoid function.
- ⁇ c The original is a hyperbolic tangent.
- Levenshtein distance is a measure of the similarity between two strings, which is referred to as the source string (s) and the target string (t) in the present disclosure.
- the distance is the number of deletions, insertions, or substitutions required to transform ‘s’ into ‘t’.
- the Levenshtein distance between two strings is given by way of illustrative expression below:
- the first element in the minimum may correspond to deletion (from a to b), the second to insertion and the third to match or mismatch, depending on whether the respective symbols are the same.
- each word is converted to a vector by using (or based on) the determined context and at least one of the set of correct sentences and the set of incorrect sentences, and the one or more complex sentences.
- the one or more hardware processors 104 assessing, using one or more recurrent neural networks, quality of the input text based on at least one of the set of generated vectors, output resulted (or extracted or derived) from the comparison between each word and the words from dictionary, the determined context and the at least one of the one or more correct sentences, one or more incorrect sentences and one or more complex sentences, or combinations thereof.
- the quality of the input text may be assessed by assigning a weightage to each of the set of generated vectors, the output resulted from the comparison between each word and the words from dictionary, the determined context and the at least one of the one or more correct sentences, the one or more incorrect sentences and the one or more complex sentences and a score for the input text is generated based on the weightage which is indicative (or characterize) the quality of the input text.
- the system 100 may generate a completeness score for the input text.
- the quality assessment made by the system 100 may be provided as a continuous feedback to the system 100 so as to learn from the pattern of the steps carried out and provide near accurate quality assessments for subsequent input texts.
- the recurrent neural networks based system 100 performs a comparison of each word from each sentence of the input text with words from a dictionary stored in a database to determine a closest recommended word for each word in the input text. For example, each word (e.g., ‘My’, ‘name’ ‘was’, and ‘John Doe’ from S1 is compared with words from dictionary to determine closest recommended word.
- S1 “My name was John Doe”
- the system 100 upon performing comparison the system 100 has determined that the closest recommended word for ‘My’ is ‘My’, ‘name’, is ‘name’ and ‘was’ is ‘is’.
- the system 100 may suggest a different word based on the level of training (or knowledge), and training data present in the database, or may simply retain or suggest the same name.
- the system 100 may then provide sentences with corrected errors as illustrated by way of example below:
- the system 100 assigns a weightage to each of the sentences and generates a score (e.g., an error correction score, say ‘6’) based on the weightage.
- a score e.g., an error correction score, say ‘6’
- the system 100 Upon comparing each word of the sentences S1, S2, S3, and S4 to words from dictionary, and identifying (or determining) the closest recommended word(s), the system 100 analyzes each of the sentences S1, S2, S3, and S4 to determine context of each word based on at least a portion of the input text. The system 100 further determines, based on the determined context, at least one of one or more correct sentences, one or more incorrect sentences, and one or more complex sentences from the input text.
- the system 100 assigns a weightage (e.g., context weightage and sentence type score) for the sentences and generates a score (e.g., context analysis and sentence type score, say ‘2’) based on the weightage.
- a weightage e.g., context weightage and sentence type score
- a score e.g., context analysis and sentence type score, say ‘2’
- the expression ‘sentence type’ refers to type of sentences such as for example, correct sentence(s), incorrect sentence(s), and/or complex sentence(s).
- the system 100 converts, each word from each of the sentences within the input text to a vector based on one or more concepts by comparing each word across the plurality of sentences of the input text to generate a set of vectors.
- the system 100 generates a ‘N’ dimensional vector representation for each word based on the concepts and by comparing each word with words within the same sentence (e.g., intra sentence) and remaining sentences (inter sentence).
- the values of that vector are generated as ‘0’ by the system 100 .
- the system 100 generates a null vector.
- the above values in each vector set depict indicate that a particular word has appeared for that many times in a particular category (which is referred to as a concept (or data chunk).
- the category (or concept or data chunk) may be stored in the database residing in the memory 102 and could be referred as trained data. For example, the word ‘beautiful’ from the sentence ‘S4’ has appeared 184 times in a concept called ‘Nature’.
- the system 100 may assign a weightage (word to vector conversion weightage), and generate a score (e.g., a vector score, say 0.21) for the vector representation.
- a weightage word to vector conversion weightage
- a score e.g., a vector score, say 0.21
- the system 100 may further provide its feedback with respect to the input text received from the user. Illustrated below is an example of the generated correct text by the system 100 :
- the above generated correct text may also be displayed in the form of corrected sentences (CS) as illustrated by way of example below:
- the system 100 generates a vector representation for the system generated text (corrected sentences—CS1, CS2, CS3 and CS4).
- the system 100 further assesses, using one or more recurrent neural networks, quality of the input text based on at least one of the set of generated vectors (vectors of the input text received from the user), the comparison, the determined context and the at least one of the one or more correct sentences, one or more incorrect sentences and one or more complex sentences.
- the system 100 assesses the quality of the input text based on the weightages assigned to output generated from steps being performed on the input text, and corresponding score being generated using the weightages (e.g., positive weightage, negative weightage, etc.).
- the system 100 performs a comparison of the generated corrected text (or generated and recommended text) and the input text received from the user. Alternatively, the system 100 performs a comparison of the vectors sets of the generated corrected text (or generated and recommended text—CS1, CS2, CS3, and CS4) and vectors sets of input text (sentence S1, S2, S3 and S4). Based on the comparison, the input text is assessed by providing a score (e.g., assessment score, say 0.47). As it can be seen that a score of 0.47 is provided to the input text depicting the quality. It is to be further noted that this score of 0.47 is due to the low coherence in the last sentence ‘S4’ which is out of the context when compared to other sentences S1, S2 and S3.
- a score e.g., assessment score, say 0.47
- the embodiments of the present disclosure may enable the system 100 to provide a completeness score to the input text based on the topic associated with the input text.
- the completeness score may be indicative of how much (portion or percentage) of the input text is complete and can be easily understood (by readers or systems).
- the embodiments of the present disclosure provide systems and methods that implement one or more recurrent neural networks (e.g., LSTM network(s)) for assessing quality of input text.
- LSTM network(s) recurrent neural networks
- the proposed system 100 determines coherence and contextual content of the input text. Further, unlike conventional assessment tools that evaluate and rate the content (e.g., essay) based on the grammatical and syntactical correctness, the proposed system 100 takes into account (or consideration) the content mapping and coherence weightage.
- the proposed system 100 assesses the input text based on plurality of parameters, for example, error correction, coherence (comparison of each word within the same sentence and across sentences), clarity and correctness, and intelligent generates and assigns a weightage to each of these parameters, and generates a score for the input text that indicates quality of the input text.
- error correction for example, error correction, coherence (comparison of each word within the same sentence and across sentences), clarity and correctness
- intelligent generates and assigns a weightage to each of these parameters, and generates a score for the input text that indicates quality of the input text.
- the system 100 is implemented with a neural network (e.g., recurrent neural network) that learns the structure and use of words in the language (e.g., English as described above).
- the language character learning may happen (or happens) based on the (large) dataset training.
- the system 100 may then predict probability of next word, being a particular entry in the dictionary, given the previous word set in a particular sentence.
- This prediction technique that is used to determine successive words based on preceded words (or context) is used to rate (or provide weightage and/or score) to the coherence of following word(s) in sentence(s) of the input text.
- the system 100 When a word is identified as an incorrect word, the system 100 performs error correction (e.g., spell and grammar check) by mapping to the nearest or closest correct word, thus making itself aware of the content or the input text. Unlike traditional assessment tools which learn from annotated corpus, the proposed system 100 learns from the written input text received from the user which makes training easier (wherein any amount of data may be scraped for the same).
- error correction e.g., spell and grammar check
- the hardware device can be any kind of device which can be programmed including e.g. any kind of computer like a server or a personal computer, or the like, or any combination thereof.
- the device may also include means which could be e.g. hardware means like e.g. an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a combination of hardware and software means, e.g.
- ASIC application-specific integrated circuit
- FPGA field-programmable gate array
- the means can include both hardware means and software means.
- the method embodiments described herein could be implemented in hardware and software.
- the device may also include software means.
- the embodiments may be implemented on different hardware devices, e.g. using a plurality of CPUs.
- the embodiments herein can comprise hardware and software elements.
- the embodiments that are implemented in software include but are not limited to, firmware, resident software, microcode, etc.
- the functions performed by various modules described herein may be implemented in other modules or combinations of other modules.
- a computer-usable or computer readable medium can be any apparatus that can comprise, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device.
- a computer-readable storage medium refers to any type of physical memory on which information or data readable by a processor may be stored.
- a computer-readable storage medium may store instructions for execution by one or more processors, including instructions for causing the processor(s) to perform steps or stages consistent with the embodiments described herein.
- the term “computer-readable medium” should be understood to include tangible items and exclude carrier waves and transient signals, i.e., be non-transitory. Examples include random access memory (RAM), read-only memory (ROM), volatile memory, nonvolatile memory, hard drives, CD ROMs, DVDs, BLU-RAYs, flash drives, disks, and any other known physical storage media.
Abstract
Description
f t=σg(W f x t +U f h t-1 +b f)
i t=σg(W i x t +U i h t-1 +b i)
o t=σg(W o x t +U o h t-1 +b o)
c t =f t o c t-1 +i t oσc(W c x t +U c h t-1 +b c)
h t =o t oσh(c t)
-
- ft: Forget gate vector. Weight of remembering old information.
- it: Input gate vector. Weight of acquiring new information.
- ot: Output gate vector. Output candidate.
σ(x)=11+e−xσ(x)=11+e−x
This combines with the normal Euclidean distance wise clustering distance ((x, y), (a, b))=√(x−a)2+(y−b)2, where x and y are coordinates, and the ordered pair (a, b) is reference point.
Where 1(a
-
- “My name was John Doe. I am fascinated by image and language comprehension in intelligent networks. I aspire to understood the functioning of evolving networks and in the process figure in the very nature of intelligence and its manifestation in beings. Birds are beautiful”
-
- S1: My name was John Doe
- S2: I am fascinated by image and language comprehension in intelligent networks
- S3: I aspire to understood the functioning of evolving networks and in the process figure in the very nature of intelligence and its manifestation in beings
- S4: Birds are beautiful
-
- S1: My name was John Doe.
- S2: I am fascinated by image and language comprehension in intelligent networks.
- S3: I aspire to understood the functioning of evolving networks and in the process figure in the very nature of intelligence and its manifestation in beings.
- S4: Birds are beautiful.
-
- S1: My name is John Doe.
- S2: I am fascinated by image and language comprehension in intelligent networks.
- S3: I aspire to understand the functioning of evolving networks and in the process figure out the very nature of intelligence and its manifestation in beings.
- S4: Birds are beautiful.
-
- “My name is John Doe. I am fascinated by image and language comprehension in intelligent networks. I aspire to understand the functioning of evolving networks and in the process figure out the very nature of intelligence and its manifestation in beings. Birds are beautiful.”
-
- CS1: My name is John Doe.
- CS2: I am fascinated by image and language comprehension in intelligent networks.
- CS3: I aspire to understand the functioning of evolving networks and in the process figure out the very nature of intelligence and its manifestation in beings.
- CS4: Birds are beautiful.
Claims (6)
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
IN201721004117 | 2017-02-04 | ||
IN201721004117 | 2017-02-04 |
Publications (2)
Publication Number | Publication Date |
---|---|
US20180225274A1 US20180225274A1 (en) | 2018-08-09 |
US11151130B2 true US11151130B2 (en) | 2021-10-19 |
Family
ID=60186080
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
US15/793,281 Active 2039-07-01 US11151130B2 (en) | 2017-02-04 | 2017-10-25 | Systems and methods for assessing quality of input text using recurrent neural networks |
Country Status (2)
Country | Link |
---|---|
US (1) | US11151130B2 (en) |
EP (1) | EP3358471A1 (en) |
Families Citing this family (68)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US9318108B2 (en) | 2010-01-18 | 2016-04-19 | Apple Inc. | Intelligent automated assistant |
US8977255B2 (en) | 2007-04-03 | 2015-03-10 | Apple Inc. | Method and system for operating a multi-function portable electronic device using voice-activation |
US8676904B2 (en) | 2008-10-02 | 2014-03-18 | Apple Inc. | Electronic devices with voice command and contextual data processing capabilities |
US20120311585A1 (en) | 2011-06-03 | 2012-12-06 | Apple Inc. | Organizing task items that represent tasks to perform |
US10417037B2 (en) | 2012-05-15 | 2019-09-17 | Apple Inc. | Systems and methods for integrating third party services with a digital assistant |
US10199051B2 (en) | 2013-02-07 | 2019-02-05 | Apple Inc. | Voice trigger for a digital assistant |
US10652394B2 (en) | 2013-03-14 | 2020-05-12 | Apple Inc. | System and method for processing voicemail |
US10748529B1 (en) | 2013-03-15 | 2020-08-18 | Apple Inc. | Voice activated device for use with a voice-based digital assistant |
US10176167B2 (en) | 2013-06-09 | 2019-01-08 | Apple Inc. | System and method for inferring user intent from speech inputs |
US9715875B2 (en) | 2014-05-30 | 2017-07-25 | Apple Inc. | Reducing the need for manual start/end-pointing and trigger phrases |
EP3480811A1 (en) | 2014-05-30 | 2019-05-08 | Apple Inc. | Multi-command single utterance input method |
US10170123B2 (en) | 2014-05-30 | 2019-01-01 | Apple Inc. | Intelligent assistant for home automation |
US9338493B2 (en) | 2014-06-30 | 2016-05-10 | Apple Inc. | Intelligent automated assistant for TV user interactions |
US9886953B2 (en) | 2015-03-08 | 2018-02-06 | Apple Inc. | Virtual assistant activation |
US10200824B2 (en) | 2015-05-27 | 2019-02-05 | Apple Inc. | Systems and methods for proactively identifying and surfacing relevant content on a touch-sensitive device |
US20160378747A1 (en) | 2015-06-29 | 2016-12-29 | Apple Inc. | Virtual assistant for media playback |
US10671428B2 (en) | 2015-09-08 | 2020-06-02 | Apple Inc. | Distributed personal assistant |
US10740384B2 (en) | 2015-09-08 | 2020-08-11 | Apple Inc. | Intelligent automated assistant for media search and playback |
US10331312B2 (en) | 2015-09-08 | 2019-06-25 | Apple Inc. | Intelligent automated assistant in a media environment |
US10747498B2 (en) | 2015-09-08 | 2020-08-18 | Apple Inc. | Zero latency digital assistant |
US10691473B2 (en) | 2015-11-06 | 2020-06-23 | Apple Inc. | Intelligent automated assistant in a messaging environment |
US10956666B2 (en) | 2015-11-09 | 2021-03-23 | Apple Inc. | Unconventional virtual assistant interactions |
US10223066B2 (en) | 2015-12-23 | 2019-03-05 | Apple Inc. | Proactive assistance based on dialog communication between devices |
US10586535B2 (en) | 2016-06-10 | 2020-03-10 | Apple Inc. | Intelligent digital assistant in a multi-tasking environment |
DK179415B1 (en) | 2016-06-11 | 2018-06-14 | Apple Inc | Intelligent device arbitration and control |
DK201670540A1 (en) | 2016-06-11 | 2018-01-08 | Apple Inc | Application integration with a digital assistant |
US10726832B2 (en) | 2017-05-11 | 2020-07-28 | Apple Inc. | Maintaining privacy of personal information |
DK180048B1 (en) | 2017-05-11 | 2020-02-04 | Apple Inc. | MAINTAINING THE DATA PROTECTION OF PERSONAL INFORMATION |
DK179496B1 (en) | 2017-05-12 | 2019-01-15 | Apple Inc. | USER-SPECIFIC Acoustic Models |
DK201770428A1 (en) | 2017-05-12 | 2019-02-18 | Apple Inc. | Low-latency intelligent automated assistant |
DK179745B1 (en) | 2017-05-12 | 2019-05-01 | Apple Inc. | SYNCHRONIZATION AND TASK DELEGATION OF A DIGITAL ASSISTANT |
US20180336892A1 (en) | 2017-05-16 | 2018-11-22 | Apple Inc. | Detecting a trigger of a digital assistant |
US20180336275A1 (en) | 2017-05-16 | 2018-11-22 | Apple Inc. | Intelligent automated assistant for media exploration |
US10818288B2 (en) | 2018-03-26 | 2020-10-27 | Apple Inc. | Natural assistant interaction |
US11145294B2 (en) | 2018-05-07 | 2021-10-12 | Apple Inc. | Intelligent automated assistant for delivering content from user experiences |
US10928918B2 (en) | 2018-05-07 | 2021-02-23 | Apple Inc. | Raise to speak |
DK179822B1 (en) | 2018-06-01 | 2019-07-12 | Apple Inc. | Voice interaction at a primary device to access call functionality of a companion device |
US11386266B2 (en) * | 2018-06-01 | 2022-07-12 | Apple Inc. | Text correction |
DK180639B1 (en) | 2018-06-01 | 2021-11-04 | Apple Inc | DISABILITY OF ATTENTION-ATTENTIVE VIRTUAL ASSISTANT |
US10892996B2 (en) | 2018-06-01 | 2021-01-12 | Apple Inc. | Variable latency device coordination |
US20200019632A1 (en) * | 2018-07-11 | 2020-01-16 | Home Depot Product Authority, Llc | Presentation of related and corrected queries for a search engine |
WO2020068877A1 (en) * | 2018-09-24 | 2020-04-02 | Michelle Archuleta | Reinforcement learning approach to approximate a mental map of formal logic |
US11462215B2 (en) | 2018-09-28 | 2022-10-04 | Apple Inc. | Multi-modal inputs for voice commands |
CA3060811A1 (en) * | 2018-10-31 | 2020-04-30 | Royal Bank Of Canada | System and method for cross-domain transferable neural coherence model |
KR20200098068A (en) * | 2019-02-11 | 2020-08-20 | 삼성전자주식회사 | Method for recommending word and apparatus thereof |
US11348573B2 (en) | 2019-03-18 | 2022-05-31 | Apple Inc. | Multimodality in digital assistant systems |
US11080545B2 (en) | 2019-04-25 | 2021-08-03 | International Business Machines Corporation | Optical character recognition support system |
DK201970509A1 (en) | 2019-05-06 | 2021-01-15 | Apple Inc | Spoken notifications |
US11307752B2 (en) | 2019-05-06 | 2022-04-19 | Apple Inc. | User configurable task triggers |
US11140099B2 (en) | 2019-05-21 | 2021-10-05 | Apple Inc. | Providing message response suggestions |
US11748571B1 (en) * | 2019-05-21 | 2023-09-05 | Educational Testing Service | Text segmentation with two-level transformer and auxiliary coherence modeling |
SG10201904825XA (en) | 2019-05-28 | 2019-10-30 | Alibaba Group Holding Ltd | Automatic optical character recognition (ocr) correction |
DK180129B1 (en) | 2019-05-31 | 2020-06-02 | Apple Inc. | User activity shortcut suggestions |
DK201970510A1 (en) | 2019-05-31 | 2021-02-11 | Apple Inc | Voice identification in digital assistant systems |
US11468890B2 (en) | 2019-06-01 | 2022-10-11 | Apple Inc. | Methods and user interfaces for voice-based control of electronic devices |
EP4010839A4 (en) * | 2019-08-05 | 2023-10-11 | AI21 Labs | Systems and methods of controllable natural language generation |
CN110705262B (en) * | 2019-09-06 | 2023-08-29 | 宁波市科技园区明天医网科技有限公司 | Improved intelligent error correction method applied to medical technology inspection report |
CN110750979B (en) * | 2019-10-17 | 2023-07-25 | 科大讯飞股份有限公司 | Method for determining continuity of chapters and detection device |
CN111061867B (en) * | 2019-10-29 | 2022-10-25 | 平安科技(深圳)有限公司 | Text generation method, equipment, storage medium and device based on quality perception |
US11544458B2 (en) * | 2020-01-17 | 2023-01-03 | Apple Inc. | Automatic grammar detection and correction |
US11061543B1 (en) | 2020-05-11 | 2021-07-13 | Apple Inc. | Providing relevant data items based on context |
US11043220B1 (en) | 2020-05-11 | 2021-06-22 | Apple Inc. | Digital assistant hardware abstraction |
US11755276B2 (en) | 2020-05-12 | 2023-09-12 | Apple Inc. | Reducing description length based on confidence |
US11490204B2 (en) | 2020-07-20 | 2022-11-01 | Apple Inc. | Multi-device audio adjustment coordination |
US11438683B2 (en) | 2020-07-21 | 2022-09-06 | Apple Inc. | User identification using headphones |
CN113743050B (en) * | 2021-09-07 | 2023-11-24 | 平安科技(深圳)有限公司 | Article layout evaluation method, apparatus, electronic device and storage medium |
CN114610852B (en) * | 2022-05-10 | 2022-09-13 | 天津大学 | Course learning-based fine-grained Chinese syntax analysis method and device |
CN116704513B (en) * | 2023-08-04 | 2023-12-15 | 深圳思谋信息科技有限公司 | Text quality detection method, device, computer equipment and storage medium |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7720675B2 (en) | 2003-10-27 | 2010-05-18 | Educational Testing Service | Method and system for determining text coherence |
US20100257478A1 (en) * | 1999-05-27 | 2010-10-07 | Longe Michael R | Virtual keyboard system with automatic correction |
US7835902B2 (en) | 2004-10-20 | 2010-11-16 | Microsoft Corporation | Technique for document editorial quality assessment |
US8577898B2 (en) | 2010-11-24 | 2013-11-05 | King Abdulaziz City For Science And Technology | System and method for rating a written document |
US20150293972A1 (en) * | 2012-12-31 | 2015-10-15 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method and device used for providing input candidate items corresponding to an input character string |
US20180046619A1 (en) * | 2016-08-09 | 2018-02-15 | Panasonic Intellectual Property Management Co., Ltd. | Method for controlling identification and identification control apparatus |
US20180300054A1 (en) * | 2015-12-28 | 2018-10-18 | Alps Electric Co., Ltd. | Handwriting input device and information input method |
US20190332670A1 (en) * | 2014-01-28 | 2019-10-31 | Somol Zorzin Gmbh | Method for Automatically Detecting Meaning and Measuring the Univocality of Text |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6424983B1 (en) * | 1998-05-26 | 2002-07-23 | Global Information Research And Technologies, Llc | Spelling and grammar checking system |
US10115055B2 (en) * | 2015-05-26 | 2018-10-30 | Booking.Com B.V. | Systems methods circuits and associated computer executable code for deep learning based natural language understanding |
-
2017
- 2017-10-25 EP EP17198271.3A patent/EP3358471A1/en active Pending
- 2017-10-25 US US15/793,281 patent/US11151130B2/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20100257478A1 (en) * | 1999-05-27 | 2010-10-07 | Longe Michael R | Virtual keyboard system with automatic correction |
US7720675B2 (en) | 2003-10-27 | 2010-05-18 | Educational Testing Service | Method and system for determining text coherence |
US7835902B2 (en) | 2004-10-20 | 2010-11-16 | Microsoft Corporation | Technique for document editorial quality assessment |
US8577898B2 (en) | 2010-11-24 | 2013-11-05 | King Abdulaziz City For Science And Technology | System and method for rating a written document |
US20150293972A1 (en) * | 2012-12-31 | 2015-10-15 | Baidu Online Network Technology (Beijing) Co., Ltd. | Method and device used for providing input candidate items corresponding to an input character string |
US20190332670A1 (en) * | 2014-01-28 | 2019-10-31 | Somol Zorzin Gmbh | Method for Automatically Detecting Meaning and Measuring the Univocality of Text |
US20180300054A1 (en) * | 2015-12-28 | 2018-10-18 | Alps Electric Co., Ltd. | Handwriting input device and information input method |
US20180046619A1 (en) * | 2016-08-09 | 2018-02-15 | Panasonic Intellectual Property Management Co., Ltd. | Method for controlling identification and identification control apparatus |
Also Published As
Publication number | Publication date |
---|---|
EP3358471A1 (en) | 2018-08-08 |
US20180225274A1 (en) | 2018-08-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11151130B2 (en) | Systems and methods for assessing quality of input text using recurrent neural networks | |
CN110765763B (en) | Error correction method and device for voice recognition text, computer equipment and storage medium | |
US10915788B2 (en) | Optical character recognition using end-to-end deep learning | |
AU2019219746A1 (en) | Artificial intelligence based corpus enrichment for knowledge population and query response | |
AU2018279013B2 (en) | Method and system for extraction of relevant sections from plurality of documents | |
US20230059494A1 (en) | Semantic map generation from natural-language text documents | |
US20220019734A1 (en) | Method and system for visio-linguistic understanding using contextual language model reasoners | |
CN110413961B (en) | Method and device for text scoring based on classification model and computer equipment | |
EP3594862A1 (en) | Resolving abstract anaphoric references in conversational systems using hierarchically stacked neural networks | |
US9734234B2 (en) | System and method for rectifying a typographical error in a text file | |
TWI567569B (en) | Natural language processing systems, natural language processing methods, and natural language processing programs | |
US11816913B2 (en) | Methods and systems for extracting information from document images | |
US20220335335A1 (en) | Method and system for identifying mislabeled data samples using adversarial attacks | |
US11675966B2 (en) | Table of contents generation | |
EP3493081A1 (en) | Systems and methods for segmenting interactive session text | |
CN107783958B (en) | Target statement identification method and device | |
CN112307749A (en) | Text error detection method and device, computer equipment and storage medium | |
JP6568968B2 (en) | Document review device and program | |
CN111428475A (en) | Word segmentation word bank construction method, word segmentation method, device and storage medium | |
US20230385659A1 (en) | Systems and methods for rules-based mapping of answer scripts to markers | |
Singha et al. | Bengali Text Summarization with Attention-Based Deep Learning | |
US20230046539A1 (en) | Method and system to align quantitative and qualitative statistical information in documents | |
US20220171925A1 (en) | Method and system for text cleaning. | |
US20230168989A1 (en) | BUSINESS LANGUAGE PROCESSING USING LoQoS AND rb-LSTM | |
CN114254627A (en) | Text error correction method, device, equipment and readable storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
FEPP | Fee payment procedure |
Free format text: ENTITY STATUS SET TO UNDISCOUNTED (ORIGINAL EVENT CODE: BIG.); ENTITY STATUS OF PATENT OWNER: LARGE ENTITY |
|
AS | Assignment |
Owner name: TATA CONSULTANCY SERVICES LIMITED, INDIA Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:TOMMY, ROBIN;SIVAPRASAD, SARATH;REEL/FRAME:043961/0798 Effective date: 20170103 |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NON FINAL ACTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: RESPONSE TO NON-FINAL OFFICE ACTION ENTERED AND FORWARDED TO EXAMINER |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: RESPONSE AFTER FINAL ACTION FORWARDED TO EXAMINER |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: ADVISORY ACTION MAILED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: DOCKETED NEW CASE - READY FOR EXAMINATION |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NOTICE OF ALLOWANCE MAILED -- APPLICATION RECEIVED IN OFFICE OF PUBLICATIONS |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: AWAITING TC RESP., ISSUE FEE NOT PAID |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: NOTICE OF ALLOWANCE MAILED -- APPLICATION RECEIVED IN OFFICE OF PUBLICATIONS |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: PUBLICATIONS -- ISSUE FEE PAYMENT RECEIVED |
|
STPP | Information on status: patent application and granting procedure in general |
Free format text: PUBLICATIONS -- ISSUE FEE PAYMENT VERIFIED |
|
STCF | Information on status: patent grant |
Free format text: PATENTED CASE |